LgbTrainConfig

LgbTrainConfig

Configuration for LightGBM training parameters.

This class encapsulates all parameters that can be passed to lgb.train(), except for validation data-related parameters. Validation data handling is managed separately: SingleModelContainer decides whether to use validation data in its train method, while CvModelContainer automatically handles it during cross-validation.

Parameters:

Name	Type	Description	Default
params	dict[str, Any]	Parameters for training.	required
num_boost_round	int	Number of boosting iterations.	100
valid_names	list[str] | None	Names of valid_sets.	None
feval	callable or list of callable	Customized evaluation function. Each evaluation function should accept two parameters: preds, eval_data, and return (eval_name, eval_result, is_higher_better) or list of such tuples.	required
init_model	(str, Path, Booster or None)	Filename of LightGBM model or Booster instance used for continue training.	required
keep_training_booster	bool	Whether the returned Booster will be used to keep training. If False, the returned value will be converted into _InnerPredictor before returning.	False
callbacks	list of callable	List of callback functions that are applied at each iteration.	required

See Also

lightgbm.train : The underlying LightGBM training function.

Examples:

>>> import lightgbm as lgb
>>> from factrainer.lightgbm import LgbTrainConfig
>>> config = LgbTrainConfig(
...     params={
...         "objective": "regression",
...         "metric": "rmse",
...         "boosting_type": "gbdt",
...         "num_leaves": 31,
...         "learning_rate": 0.05,
...     },
...     num_boost_round=100,
...     callbacks=[lgb.early_stopping(10), lgb.log_evaluation(50)],
... )

Attributes

params instance-attribute

params: dict[str, Any]

num_boost_round class-attribute instance-attribute

num_boost_round: int = 100

valid_names class-attribute instance-attribute

valid_names: list[str] | None = None

feval class-attribute instance-attribute

feval: (
    _LGBM_CustomMetricFunction
    | list[_LGBM_CustomMetricFunction]
    | None
) = None

init_model class-attribute instance-attribute

init_model: str | Path | Booster | None = None

keep_training_booster class-attribute instance-attribute

keep_training_booster: bool = False

callbacks class-attribute instance-attribute

callbacks: list[Callable[..., Any]] | None = None